Study And Design On Aircraft Fault Diagnosis Expert System Based On Case-based Reasoning And Fault Tree Analysis Integrated

This paper studies the background of existing aircraft fault diagnosis technologyand aircraft faults characteristics, puts forward a new idea of Case-Based Reasoningand fault tree analysis integrated for aircraft fault diagnosis expert system.The paper firstly analysed structure of Case-Based Reasoning (CBR), usedframe-based representation method to store the historical cases, and retrieved cases bynew algorithm combines Character field matching with nearest neighboring algorithmwhich makes fault location more accuracy. The superiority of the algorithm wasverified by two actual cases and get good results. Besides, the case reuse, case reviseand case retain had been all designed in this paper. Secondly, the paper depth studiedthe fault tree analysis, adopted object-oriented knowledge representation when builtthe fault tree knowledge database, detail described how to built fault tree andmathematical basis of it.Knowledge database and system inference engine in fault diagnosis expert systemhad been detail designed in this paper. Fault tree knowledge and historical casesknowledge were stored in different file formats in the knowledge database. First of alldiagnosed fault by the fast and efficient Case-Based Reasoning with existingphenomenon, if the fault was successful located, stopped diagnosing. Otherwise turn toFault Tree Analysis for accurate diagnosis. Case-based reasoning diagnosis and faulttree diagnosis would call different file formats at runtime, thus ensured a reasonableuse of the database and greatly reduce the time of diagnosis. Whether to keep new casein case database depended on the similarity of original case with new case. Invalid oldcase should be deleted termly. The fault unable be solved by Fault Tree Analysis wouldbe stored in the case database and fault tree database after being solved.Maintenance subsystem of the expert system could delete, increase and modifythe principle data in the system. The subsystem ensured reasonable validity of the knowledge stored in the expert system knowledge database. Most of the actualoperation interfaces of the system had been shown in the last diagram, the systemworked in a good condition after the initial trial and basically reached the standard ofputting into market.